DeinFreund

Their algorithm is based on splitting the embeddings into a bit-wise representation and then sampling each bit based on the secret key, preserving the same likelihood distribution as with random sampling. (Given that the key is random)

They say this works wlog for more complex embeddings, by encoding each token as a bit string. Could someone explain this generalization to me?

If we have 4 tokens, 00, 01, 10, 11 with probabilities 0.5 for 00 and 11 and probability 0 for 01 and 10. Going through bit by bit, how will the algorithm guarantee not to produce 01 or 10?

Thanks for the correction! Was probably my local bias tending to either cheaper reflectors or using public observatories.

I'd say the most important factors for astrophotography are in order:

1. Location and weather

2. Optics

3. Sensor, tracking and post processing

They can't control the first, but it seems they put everything in (3) instead of (2). The telescope is a 15cm refractor. I don't know any amateur astronomer using a refractor, let alone one with this budget. Refractors introduce further optical aberrations compared to first surface mirrors, usually collect less light and have shorter focal lengths for the same telescope size. Is there something I'm missing?